News Qualcomm CEO says Arm taking 50% of the Windows PC market in five years is realistic — some OEMs already expect Snapdragon chips to be 60% of their...

If laptops containing these ARM SoCs can be made cheaply I could see it, non-US/EU countries would be likely to snap them up at the right price point. However, at price points we've seen to date, I have a hard time believing 50% is possible in 5 years.
 
It is very much not going to happen. PCs are used completely differently to tablets and phones...

People expect the best CPU for their desktops and laptops. Also with full backward compatibility.

They are dreaming...
 
Qualcomm isn't a household name, average PC consumers barely know AMD exists. The average buyer will continue to purchase based on brand and price point, Qualcomm has neither. Businesses will upgrade based on compatibility and price point, Qualcomm has neither. The ONLY way they etch out a piece of the CPU market pie is by coming in cheap and good, something they've never been particularly good at either.
 
It is very much not going to happen. PCs are used completely differently to tablets and phones...

People expect the best CPU for their desktops and laptops. Also with full backward compatibility.

They are dreaming...
If they do what I've been saying for years: Add an ARM based PCI-e card (co-processor) for combability with ARM based Windows and Android programs. Think of this is a card similar to a GPU and resides in an X16 slot right next to it. Your OS could run on the ARM card (windows options) while your x86-64 cpu is freed up to run other programs.
 
he is crazy.

a lot of the market refuses to even update windows and he assumes they can make em not only upgrade but also switch to a new unfamiliar thing?

ARM will take over x86 but its not goign to be nearly as fast as this guy thinks.

ARM wont overtake a majority until most devs back it w/ same level(or more) that they give x86.

More people care about "it just works" than any slight performance gains.
 
  • Like
Reactions: iLoveThe80s
PCs are used completely differently to tablets and phones
yes, but ARM based cpus for pc's isnt anything new.
amd did it long long ago and iirc intel's entire x-scale is arm based (and they had that since early 2000's)

Apple's M series chips are proof of how good ARM cpu's can be in powerful stuff (i can hate apple and admit the product itself is great) beyond just phones and tablets.
 
Qualcomm isn't a household name, average PC consumers barely know AMD exists. The average buyer will continue to purchase based on brand and price point, Qualcomm has neither. Businesses will upgrade based on compatibility and price point, Qualcomm has neither. The ONLY way they etch out a piece of the CPU market pie is by coming in cheap and good, something they've never been particularly good at either.
Well Qualcomm does have brand recognition, but it's mainly in telecomm.
However, that overlaps with mobile computing, which means they aren't a complete unknown to the average person, and the best time to be a new entrant is when everything is in a major upheaval in a marketplace that was stagnant until recently, and whose potential upside potential has expanded dramatically.

The business segment for Ai PCs is still questionable as security & utility evolves. The top use-case of the Enterprise space [workstations, servers] is already there and buying and churning, but it's the big middle of the business PC market that can see some desire to move from coupke of year old Core i5 running browser apps and Excel, etc. To something with a 'unique feature' to the thin-client segment.

Heck, even the big boys don't know how to market this, they're still advertising video conferencing tweaks and calendar tweaks as the MUST HAVE Ai features?

It's pretty easy to justify/sell ROI for things like FTE changes (ex. content/spec review/generation , RFQ/P/I analysis/prep, etc). That could save enough man hours to justify most Ai PCs multiple times over. Just identifying minimum specs/requirements for proposals let alone writing up how you tick every box used to take weeks for dedicated proposal/marketing teams. This can now be done by a couple of people in a few days (with heavy proof-reading for errors/hallucinations).

Right now is the best time for a company to take a risk to get early adopters in the SMB space & consumer segment, especially for those whose idea of Ai is limited to their Samaung / Google phones, where Qualcomm has brand recognition.
Also, it's more likely the front end of these Ai PCs ie OEM badges (ASUS, DELL, HP, LENOVO) and software (Micro$oft, CRMs, etc) will matter more to most consumers and businesses than if their Ai PC has an AMD, intel, Qc, nV or the myriad of other ARM players in it, AS LONG AS IT CAN DO THE JOB FOR $X per Satisfaction Unit.

Qualcomm's pricing will also improve over time as well, heck the value/perf vs their earlier C series is already taking it from disruptor-only purchases to slightly wider adoption. It's likely to snowball from there.

We've already seen this bifurcation/ un-coupling in mobility outside of the the very top where Exynos vs Snapdragon still maters. But for the mid-range if a Mediatek can get most people more features for less dollars and no additional risk, then that premier chip/IHV holds less value. Of course Apple is the odd-ball exception to this rule especially in N.Am.

It's coming to PCs, just not the high end market which most folks here deal with, vs the volume purchasers (consumer p/busines) who will just want to get a task done, and likely will use aesthetics or ergonomics as their deciding factor instead of CPU/GPU/NPU mfr.

Now is Qualcomm's best time to help push the market beyond the X86 environment, with their depth of experience in ARM eArch they aren't just taking a stab at it like many others, and with M$' obvious support is changing the balance of power with AMD & intel, there is a lot of potential ther for Qualcomm, The only question is if they can convert that potential to significant gains, not just have a great part (like the nV Tegra for example) and then return to the fringes.

It's also hard to say any market swing is impossible considering the large and opposing swings in AMD's fortunes in CPUs vs GPUs and their previous positions. If everything falls into place 50% is quite doable, especially if AMD, intel and nVidia get distracted by the HUGE revenue/profits of the Enterprise segment that makes personal computing a money loser for their engineering focus.
 
he is crazy.

a lot of the market refuses to even update windows and he assumes they can make em not only upgrade but also switch to a new unfamiliar thing?

ARM will take over x86 but its not goign to be nearly as fast as this guy thinks.

ARM wont overtake a majority until most devs back it w/ same level(or more) that they give x86.

More people care about "it just works" than any slight performance gains.

Even more of the market hasn't turned on their 'PC' in years and have already moved to other devices/platforms. The one are that could bring many of them back for NEW MONEY purchases is a scenario where we can take those new Ai feature you kinda played with via ChatGPT on your phone/tablet or editing your pictures, and showing you how much better it can be on an Ai PC that cost less than your latest high end phone / tablet.

The real money isn't in the current traditional PC users who do know the major players, that's a shrinking, less profitable market being undermined by the huge money in the Enterprise space. The real money/volume is if they can market and sell the idea of people moving back from their phones/tablets to a PC for features that can be boosted on PC.

Granted, I think that that's the WAY overly optimistic view, as I think the low-res/low-quality short-attention-span generation really won't go the PC route, except those already there (like content creators).

People expect the best CPU for their desktops and laptops. Also with full backward compatibility.

They are dreaming...

Huh? Nah, I doubt that.

WE expect the best CPUs, GPUs, RAM, SSDs, etc... 'people' just want a box that does a task and looks/feels nice doing it for a 'good price'.

Your argument is similar to the audiophiles quoting SACD or DSD now, while the majority and the money is in the plebes buying $50-150 headphones for their MP3s or cheap subscription music (because premium HD audio costs money).

Backwards compatible? For some, but usually for the cheaper consumers who aren't prone to disposable tech. Also, having cross-platform services like 365 and Adobe Cc, etc means the backwards compatible thing can be focused elsewhere too.

The number of apps I have that no longer work in M$ compatibility mode or have been sundowned by Apple & Google is about as many as I currently run.

Subscription services suck for many (especially me) but it has made it easier for people to just ignore compatibility on their hardware end and have the software folks deal with it on their end.

I think most people are arguing the enthusiast market, when much of this will play out in the general public low-mid end, who don't care much under the hood as the logo on the cover, like most Apple users didn't care about intel vs Arm, just as long as it was their brand, and few people cared Google Pixel with Qualcomm or Tensor. As long as it looks/feel OK and is reliable, they likely won't ever know.

That's just my two frames worth, your mileage may vary, 🤷🏻‍♂️
 
Intel is saying Skymont has +2% IPC vs Raptor Cove. While consuming much less power. Apparently a big upgrade vs crestmont and gracemont.

Apparently these aren’t your grandfather’s atom cores. If true, then new atom cores like darkmont will be even more performant. So Intel might may well fend off the ARM invasion.

The next couple of years will be extremely interesting as the vendors are staking out their battle positions. And who knows who else will enter the market? Might mediatek throw its hat in the ring? What about nvidia…. If they really wanted to they could make a killer APU with ARM + RTX graphics.

We haven’t seen competition like this since the 90s when multiple vendors were trying to copy intel’s x86 nomenclature and Intel had to move to using the pentium name since numbers can’t be trademarked.
 
Personally I believe the hype on this. Mac M chip line proves that this can well work extremely well as long as the end user cant tell and same performance with better battery life itll happen for laptops for sure.
 
WE expect the best CPUs, GPUs, RAM, SSDs, etc... 'people' just want a box that does a task and looks/feels nice doing it for a 'good price'.

Your argument is similar to the audiophiles quoting SACD or DSD now, while the majority and the money is in the plebes buying $50-150 headphones for their MP3s or cheap subscription music (because premium HD audio costs money).
True, but most of those groups of people aren't the sort to argue about it on forums... they too busy living to do that. Buy something on a whim cause it looks good... replace it if it breaks... must keep consuming.

People who want better CPU or Hi Res Audio are the ones who will quibble over little details. Don't get me started on ANC headphones... :)
 
  • Like
Reactions: KnightShadey
if Intel Lunar Lake matches Qualcomm battery life, Elite X chips are dead on arrival.

Intel and AMD beats Qualcomm in both AI and GPU performance.

Keep in mind that windows for ARM has a translation layer like apple rosetta. The performance is simply not there to match Lunar lake and Zen 5.
 
if Intel Lunar Lake matches Qualcomm battery life, Elite X chips are dead on arrival.

Intel and AMD beats Qualcomm in both AI and GPU performance.

Keep in mind that windows for ARM has a translation layer like apple rosetta. The performance is simply not there to match Lunar lake and Zen 5.
Intel and AMD have the advantage when there is no power constraints. So on desktops, I feel it’s hard to challenge them. But in a power constraint situation, the tide turns. Perhaps LNL is going to be a lot of power efficient, but I think it’s worth waiting to see the end results than trying to conclude using some skewed Intel marketing slides. I feel it is not impossible for Intel to make more efficient processors. At idle, you can get the efficient cores to take over, and that’s what the efficient core should be used for. But I am more interested under sustained load, because that’s where you will need to balance power draw vs performance. And this is where I feel ARM SOCs have the upper hand.

The concern around translation layer may be a problem for the first couple of generations because as soon as the adoption rate improves, there’s a strong case to optimize apps for x64 processors. But I agree in the near term, performance penalty to the translation layer is unavoidable. Qualcomm may have MS support here, but unlike Apple that’s in control of both the hard and software, I feel there will be more hiccups for early adopters.
 
  • Like
Reactions: jlake3
My feeling is that most readers of this forum, do not appreciate that their opinions and decisions do not matter when it comes to the truth of this prediction.

50% market share can be made true, simply by steering those who make the purchase decisions for corporate "desktops" which are mostly laptops already.

That is very few people who are easily swayed by Microsoft's marketing, which has been incredibly successful with FUD. These decision makers are driven by CMA and bottom line.
And for the bottom line, some illusional gains are typically ok. After all you have to show gains every year.

Very little chance you'll find these people here.
 
  • Like
Reactions: KnightShadey
Huh? Nah, I doubt that.

WE expect the best CPUs, GPUs, RAM, SSDs, etc... 'people' just want a box that does a task and looks/feels nice doing it for a 'good price'.



Backwards compatible? For some, but usually for the cheaper consumers who aren't prone to disposable tech. Also, having cross-platform services like 365 and Adobe Cc, etc means the backwards compatible thing can be focused elsewhere too.
It’s anecdotal of course, but from what I’ve seen the people who are still buying a “box” that lives on/under their desk generally have something in mind for it that’s above and beyond Office and internet, and are buying a box because they are looking for more performance. They might not want “the best”, but most of them are running 60-class dGPUs or better.

And don’t discount the importance of backwards compatibility! Cheaper consumers really don’t love having the software they bought a permanent license for turned into an ongoing subscription, and lots of people have something older they want to keep and “it runs Windows” would make them expect that they could. A bad translation layer launch could sour people towards further purchases.

The ones who don’t need backwards compatibility and don’t need performance have been moving towards lower-cost iPads and Chromebooks rather than Windows laptops. And a lot of the people with fairly performant boxes on their desk bought iPads and Chromebooks for when they’re away from them where they used to own a Windows laptop as well.

And if the path to 60% of consumer PCs is rough, then they’re gonna have to absolutely dominate in Corporate and Education sales to bring up the average. Education is extremely cost sensitive and there’s no X Pro/Elite devices that strike me as “Budget” yet, so a bit early to project a win there. Businesses could probably switch over people who aren’t using specialized software, but actively working while being away from power for an extended time is a pretty rare occurrence, and it’s safe to assume Intel and AMD are going to cut prices and release new offerings rather than completely roll over and give up on corporate sales. Plus corporate IT purchases tend towards conservative, proven solutions.
 
If laptops containing these ARM SoCs can be made cheaply I could see it, non-US/EU countries would be likely to snap them up at the right price point. However, at price points we've seen to date, I have a hard time believing 50% is possible in 5 years.
It was leaked that Qualcomm is selling the Snap X at around $120 to OEMs vs $300 of Raptor Lake and $450 of Meteor Lake. Even if the user won't see all the drop, the OEMs are flocking to Qualcomm right now as the profit margin with them is hugely more, while tests and benchmarks also gives them as winners for every single task and battery duration vs M3 & Meteor Lake, even if not by a big margin.

if Intel Lunar Lake matches Qualcomm battery life, Elite X chips are dead on arrival.

Intel and AMD beats Qualcomm in both AI and GPU performance.

Keep in mind that windows for ARM has a translation layer like apple rosetta. The performance is simply not there to match Lunar lake and Zen 5.
Even if Lunar Lake beats them, the pricing difference (it will never be cheaper than Meteor Lake) is too brutal, so OEMs will avoid Intel except for the top-end laptops or to get more supply after Qualcomm caps (though the Snap X is not being in a cutting edge node, so it can be mass-produced already, but Intel can spam infinite Alder/Raptor Lake).

Also Snap X doubles the AI performance of Meteor Lake and beats Zen 4.5 by 30%, while just using 10W. It is closer in GPU performance, but all the three performs similarly: superb for integrated, but Qualcomm is doing it a fraction of the W use.

Also keep in mind Qualcomm is already beating Meteor Lake & M3 after the performance hit due the translation layer, and Qualcomm is already preparing a sucessor for next year to compete with Lunar Lake, too, though we have yet to see how much Intel ends up delivering to see if Qualcomm can beat them or not with the next Snap X.
 
Last edited:
Also Snap X doubles the AI performance of Meteor Lake and beats Zen 4.5 by 30%, while just using 10W. It is closer in GPU performance, but all the three performs similarly: superb for integrated, but Qualcomm is doing it a fraction of the W use.

Also keep in mind Qualcomm is already beating Meteor Lake & M3 after the performance hit due the translation layer, and Qualcomm is already preparing a sucessor for next year to compete with Lunar Lake, too, though we have yet to see how much Intel ends up delivering to see if Qualcomm can beat them or not with the next Snap X.
The performance figures are still not properly validated. Qualcomm is using ARM's CPU core design, they cannot magically create a higher performance cpu core operating at the same power level than Apple M3.

The GPU performance are on par with the low end 15W designs from Intel and AMD's previous gen. the latest revisions will outperform them.

https://www.tomshardware.com/tech-i...eliteplus-benchmarks-claimed-to-be-fraudulent
 
Last edited:
It’s anecdotal of course, but from what I’ve seen the people who are still buying a “box” that lives on/under their desk generally have something in mind for it that’s above and beyond Office and internet, and are buying a box because they are looking for more performance. They might not want “the best”, but most of them are running 60-class dGPUs or better.

By box, I didn't mean a case, I meant a widget, wethers that's a laptop, 2-n-1, tablet, AIO,, blade, etc.

Qualcomm is DEFINITELY not talking about the ATX crowd.
Even in the business space that represents maybe 5% of new purchases at best, replaced by thin-clients, NUCs, and laptops.

And don’t discount the importance of backwards compatibility! Cheaper consumers really don’t love having the software they bought a permanent license for turned into an ongoing subscription, and lots of people have something older they want to keep and “it runs Windows” would make them expect that they could. A bad translation layer launch could sour people towards further purchases.

I understand that, but that segment is dwindling, and with fewer and few install options that slice of the pie is smaller... even them trying to figure out how does that install disc from their Win 7/8 software get loaded on their Win 11 machine now without an optical drive? Sure knowledgeable people have lots of options, just get a cheap usb drive or make a virtual disk/ISO, but it that enough to question Qc's 50% for backwards compatibility? Doubtful.

In the corporate space it's slowly but surely becoming an as X approaches 0 because it's no longer an on-device consideration for 90+% of them, and licensing/subscription is usually the only model... especially for Windows. Even then Security beats compatibility every time, so as long as they get that right they'll be a factor.


The ones who don’t need backwards compatibility and don’t need performance have been moving towards lower-cost iPads and Chromebooks rather than Windows laptops. And a lot of the people with fairly performant boxes on their desk bought iPads and Chromebooks for when they’re away from them where they used to own a Windows laptop as well.

Again, I think some of y'all are missing the point.... gaming PC buyers still aren't the biggest Windows market, so the Chromebook, Mac, Tablet, Phone defections are all relative to pruning the tree at the leafy middle, but the thin top is gamers and workstations, the fat trunk remains the generic cheap efficient business widget (whatever the form factor).

And if the path to 60% of consumer PCs is rough, then they’re gonna have to absolutely dominate in Corporate and Education sales to bring up the average.

Education is extremely cost sensitive and there’s no X Pro/Elite devices that strike me as “Budget” yet, so a bit early to project a win there.


I think you're mixing their OEM % with their Windows Percent.... and you're got it backwards again. They could dominate the consumer PC market, and they's still have to grab almost half the Business market to reach 50% windows PCs because it's about a 3:2 split at best between the two (usually 2:1), with consumer shrinking quicker than commercial.

Education is also a fraction of either the other two, so it's really a rounding error difference to this discussion, and if anything price, efficiency would favour Qc for Education too.

lFvlUY7jgLVtMO_xn665iQuFRVf5_MY7.png


Also adding the "Ai PC" driver for sales as a major growth factor especially for Windows Copilot+ PCs that are the new 'Cool thing' just strengthen the argument against traditional roles and mfrs.

kcO7NJuZJEEYk0lAle40qPfcWKlwQ09I.png


Businesses could probably switch over people who aren’t using specialized software, but actively working while being away from power for an extended time is a pretty rare occurrence, and it’s safe to assume Intel and AMD are going to cut prices and release new offerings rather than completely roll over and give up on corporate sales. Plus corporate IT purchases tend towards conservative, proven solutions.
Most commercial PC buys now are laptops, and while many may be plugged in 24/7 at a cubical or at home, their ability to not be tied down is one of their prime selling points, not only for work on the road or from the couch at home, but for collaboration, staffing level changes, and hotswapping, etc. That they can also do XYZ tasks at lower power is another strong selling point, even if it's plugged-in 24/7 because their idle/sleep state is even lower, and energy costs DO still drive many conversations even if the old bunch of fluro lights above their cube are eating more energy than a dozen laptops.

I can speak from experience, corporate IT purchasing (and the flip side of sales) are more interested in value, then security, then all other considerations. If it runs Windows, and gets them secure remote access/control/management... then that option is considered, Warranties, QOS targets/metrics, and service agreements cover the rest.
'Proven solutions' as a focus moved from hardware to software starting about a decade ago. And, if you're large enough in the Enterprise space it's the vendor's problem, not yours, and usually with penalties for failure that make one bad choice a blip. Again servers and network hardware being the exception to the rule.

The number of people in purchasing or IT support nowadays that even know what MTBF means outside of the server environment is dwindling quickly.

Qualcomm are likely overly optimistic in their targets, but that's different from un-realistic. It's so far out that a lot of things can change, 5 years ago we were still 9 months away from COVID which dramatically changed work environments, remote work, VoIP, teleconferencing. Might Ai PC adoption be the next major shift? Only time will tell, but at least Qc has a seat at the table.
 
  • Like
Reactions: Nyara
I hope they achieve their goal with other players. I am over x86 and the duopole Intel/Amd only for the past 20 years.

We need much more competition on the cpu market.
 
I'm thinking about buying the following Lenovo IdeaPad 5G 14Q8X05 with the 8CX processor:

https://psref.lenovo.com/Detail/IdeaPad_5G_14Q8X05?M=82KF000PSA

Qualcomm® Snapdragon® 8cx (8C, 8x Kryo 495 @ 2.84GHz)

Integrated Qualcomm® Adreno™ 680 GPU

8GB Soldered LPDDR4x 1866

I can buy it for the equivalent of about US$330 brand new, sealed in the box

Not a allot of laptops at that price comes with a 1080P IPS display, backlit keyboard, IR webcam, 512GB SSD, Win 11 Pro and a 5G modem.

I'm not planning on doing heavy stuff on the laptop, and will be used for Internet browsing, watching videos, doing some MS Office work and maybe playing some older game titles that was originally designed for Windows RT, which should run natively. All the apps that I am planning of running will be native for Windows on ARM,

Or do you reckon it's a bad deal @ US$330 brand new sealed in the box? (It was listed for US$1,300 when it was introduced to market in November 2021)